Unverified50% confidenceBenchmarkExact time
For the schemagate benchmark's paired McNemar test, the p-value is about 1e-49 on Spider and about 1e-36 on BIRD.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe fine-tuned 406M model achieved 36.33 ROUGE-1 on the test set, while the authors' own 1.2B system achieved 35.41 ROUGE-1, and the two cannot be statistically distinguished.61% similarUnverifiedIn schemagate's benchmark, on the Spider dataset's 1034 questions the pure mathematical ranking accuracy is 80.3%, and 96.5% after re-ranking; on the BIRD dataset's 1534 questions the pure mathematical ranking is 88.6%, and 96.7% after re-ranking.61% similarUnverifiedThe self-developed W4 kernel's perplexity (probe ppl) is 15.76, better than bitsandbytes NF4's 18.98.59% similarUnverifiedOn the LocalLLaMA/typed-decisions leaderboard, aplomb-1 and Darwin-27B-ZTC tie for second with scores of 0.74; ProKope-421M-8K scores 0.72, ProKope-421M scores 0.71, and typed-decisions-minilm-l6-specialist scores 0.61.56% similarUnverifiedThe authors tested a variety of vision architectures ranging from D=7,850 to D=4,203,522 (about 4.2 million parameters) on the full MNIST benchmark.55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1049520API
curl https://kongchang.com/api/v1/knowledge/claims/1049520MCP
get_claim(id=1049520)